avalon sciences limited td000795 · communicated to unauthorised persons or persons outside the...

© This document is the exclusive property of Avalon Sciences and its contents must not be copied or

communicated to unauthorised persons or persons outside the company without written consent form

ASL.

Prepared by:

XBob Pay

Senior Software Engineer

Authorised by:

XAndy Maynard

Software Team Leader

Document Number: TD000795 Issue: 04 Date: 14-12-2015

UNCONTROLLED COPY Page: Cover Page

Avalon Sciences Limited

Avalon House, Somerton Business Park, Bancombe Road Somerton, Somerset, UK, TA11 6SB

AVALON SCIENCES LIMITED

TD000795

MIRF6 Specification

UNCONTROLLED COPY

This document may be used for reference but its circulation and updating is not controlled. This may not be the latest version currently in force.

MIRF6 Specification TD000795 14-12-2015 Issue 4

of 13

Contents

1 Introduction ...................................................................................................................................... 4

2 Changes introduced at version 5 .................................................... Error! Bookmark not defined.

3 Note about UTC time and time-stamp ............................................................................................ 4

4 MIRF file structure ........................................................................................................................... 4

5 Channel attributes ........................................................................................................................... 5

6 General Header Block ..................................................................................................................... 6

7 Channel Header Block .................................................................................................................... 9

8 Channel structure .......................................................................................................................... 10

9 Sample Encoding formats ............................................................................................................. 11

9.1 Format –1 (FORMAT_VARIES) ............................................................................................ 11

9.2 Format 0 (DAQ IFP) (obsolete) ............................................................................................. 11

9.3 Format 1 (Multilock IFP) (obsolete) ....................................................................................... 11

9.4 Format 2 (Geochain IFP, also used for SST500 & MSR) ..................................................... 11

9.5 Error handling ........................................................................................................................ 12

9.6 Format 3 (32-bit integer) ....................................................................................................... 12

9.7 Format 4 (32-bit floating point) .............................................................................................. 12

9.8 Format 5 (24-bit Integer) ....................................................................................................... 12

9.9 Format 6 (16-bit Integer) ....................................................................................................... 12

9.10 Format 7 (Delta IFP) (obsolete) ............................................................................................ 12

10 Sample decoding .......................................................................................................................... 13

11 Legacy MIRF exceptions ............................................................................................................... 13


of 13

Amendment record

Issue: Date: DCR ref: Prepared: Comments/Changes:

1 27-01-2014 N/A R. Pay First Issue

2 30-04-2014 N/A R. Pay Added missing channel Descriptor codes

3 12-05-2014 N/A R. Pay Added previously undocumented error control flag

4 14-12-2015 N/A C.Bird Added High Side Indicator data


of 13

1 Introduction

MIRF (Media Independent Recording Format) is designed to simplify the management of VSP and

Microseismic data recorded on a PC based data acquisition system. The present MIRF5 is a superset of the

original MIRF specification, providing for the additional requirements of the current generation of data

acquisition systems and microseismic recording.

A MIRF record is a PC compatible file, so can easily be transferred between different types of media using

standard PC file utilities.

2 Changes introduced at version 6

1 New Overlap New High Side Indicator field, for supported Receiver Types.

2 Description of the out-of-range floating point errors, when using AS-272 digitisers in formatted

mode.

3 New formatted test modes

3 Note about UTC time and time-stamp

Support for a precision time-stamp was introduced at software version 1.85. Calendar time in the header is now

UTC (universal coordinated time). Exactly how close to UTC time depends on the acquisition hardware. If

timestamp_mode = 2, then the acquisition controller is phase-locked to a GPS receiver and the timestamp is

very accurate (within 10us). This feature supports multiple site data acquisition.

4 MIRF file structure

A MIRF record (or file) consists of a variably sized header block followed by de-multiplexed data channels. The

data channels are normally encoded in the native format of the acquisition system with which they are recorded.

A format code in the header instructs translation software how to interpret the data.

All of the information in the header block is represented in 32-bit values, mainly as signed integers but also as

32-bit IEEE754 floating point where necessary. The 32-bit values are stored in the PC compatible little-endian

byte order.

Unless stated otherwise, all depths, offsets and distances are measured in the same system of units, either

millimetres or millifeet as specified by the "measurement units" field.

The header block is split into two sub-blocks. The first block is always 512 bytes in size and contains general

information which applies to all channels. The size of the second block depends upon the number of channels

defined within the record.

The first block is known as the "general header block". One of the fields in the general header block specifies

the number of channels defined in the record. This information is used to determine the size of the second block,

known as the "channel header block" and which contains an array of contiguous channel structures, requiring 64

bytes per channel.

Please note that the number of channels defined within the header is not necessarily the same as the number of

channels for which data actually exists. Each channel may have a different number of samples recorded and in

particular a channel which has zero samples (NS = 0) should be considered to be "switched off".


of 13

5 Channel attributes

The channel attributes are stored in the channel header block and provide information specific to a particular

channel.

Each channel is assigned an "owner" and possibly also a "descriptor" attribute. The owner attribute tells us to

which receiver the channel belongs and the descriptor attribute tells us the signal component.

Other attributes include dc offset and voltage scaling factors which are required for the signal decoding

algorithm.


of 13

6 General Header Block

Integer Name Description/Possible values

1 MIRF_version 6

2 File_type 0 = raw data

1 = stacked data

3 Format_code -1 = format varies (get format from channel header block)

0 = DAQ IFP (obsolete)

1 = Multilock IFP (obsolete)

2 = Geochain IFP

3 = 32-bit integer

4 = 32-bit IEEE754 floating point

5 = 24-bit integer

6 = 16-bit integer

7 = 16-bit Delta IFP (obsolete)

4 Correlation_flag 0 = raw data

1 = correlated data

5 Controller_type 10 = DAQ (obsolete)

20 = MAU (obsolete)

30 = ASP (obsolete)

40 = GSP

50 = TSP (obsolete)

60 = SIMULATOR

70 = MONITOR

6 Tool_system 0 = ANALOG

1 = MULTILOCK (obsolete)

2 = SST500 (obsolete)

3 = GEOCHAIN

4 = MSR (obsolete)

5 = SMWD (obsolete)

6 = DELTA (obsolete)

7 = RWSPSM (obsolete)

7 Channels_defined Determines size of channel header block

8 Test_mode 0 = TEST OFF (Normal)

generic test modes

-1 = IMPULSE_TEST

1 = SINEWAVE

2 = DAQ_DISTORTION_LO (obsolete)

3 = NOISE_TEST

4 = DAQ_PREGAIN (obsolete)

VSP Test Modes

-2 = PULSE_ALL

-3 = PULSE_VZ

-4 = PULSE_HX

-5 = PULSE_HY

-6 = PULSE_DH

-7 = EPULSE

-8 = TAS_SINE

-9 = SIM_SINE

-10 = TAS_ZEROES

Geochain telemetry test modes


of 13

-30 = LRX_UNEQ_LINE_RESPONSE

-31 = LRX_UNEQ_RANDOM_DATA

-32 = LRX_EQ_DATA_QUALITY

-33 = LRX_MTX_LOOPBACK_TEST

Formatted Test Modes

20=DIGI SINE BURST

21= DIGI SINE

22=SHIFT BIT

9 Dataset_id Used to group related data.

For a simple survey, dataset_id = source_id, by default. If same

source is used on multiple runs, then records from each run are

normally grouped by dataset_id.

10 Year UTC year at first sample.

11 Month UTC month at first sample.

12 Day UTC day at first sample.

13 Hour UTC hour at first sample.

14 Minute UTC minute at first sample.

15 Second UTC second at first sample.

16 Timezone_bias_seconds Number of seconds by which local time differs from UTC time.

17 Source_id Source number. Normally zero for a system test record.

18 Number_of_receivers The number of receivers is supplied by the user. It should be

interpreted as the maximum number of receivers for which data may

be present.

19 Receiver_number Obsolete – used for early single-receiver stack files.

20 Number_in_stack Number of records used to create a stack file.

21 Measurement_units 1 = millimetres

2 = millifeet

All elevation, depth and distance values in the header are

represented in these units, unless otherwise stated.

22 Receiver_polarity Receiver polarity for the tool system, as defined in the sensor

configuration file. This is currently only meaningful for a vertical

sensor.

0 = undefined

1 = SEG normal

-1 = SEG reverse

23 Source_reference_channel Multiple channels may be assigned to the REF descriptor. This field

identifies the particular channel holding the reference signal

applicable to this particular record.

24 reserved

25 Record_number Record number is unique within a survey and forms part of the

record filename. Record numbers are allocated sequentially.

26 Stack_number Records with the same stack number may be stacked with each

other.

27 Fix_number Positioning system shot fix identifier. May be used to merge missing

or revised source coordinates into a dataset. Operators often strive to

keep fix number the same as the record number, but this is neither

guaranteed nor required.

28 Tool_MD Measured depth at the tool reference. The tool reference is the point

on the tool string that is positioned at the well reference elevation at

the time that the depth measurement is zeroed. For safety reasons,

the tool reference is normally either at the top tool or at the

telemetry package.


of 13

29 SIus Sample interval, measured in microseconds. To properly interpret a

Geochain telemetry test record, the following substitutions must be

performed. For 1us, read 0.9765625us For 2us, read 1.953125us For

4us, read 3.90625us.

30 Line_number Not used.

Dataset_id is the preferred method of grouping records.

31 Gun_pressure Airgun firing pressure measured in psi.

32 Software_version_x100 The ACQ software version used to acquire the record, multiplied by

100. E.g. software version 1.85 is stored as the integer value 185.

33 SCX Source easting coordinate relative to wellhead.

34 SCY Source northing coordinate relative to wellhead.

35 TCX Target source easting coordinate relative to wellhead.

36 TCY Target source northing coordinate relative to wellhead.

Note that TCX and TCY now normally contain the same values as

SCX and SCY, so may be ignored.

37 WRE (aka WREF) Well reference elevation, also known as “logging datum” or

“drilling datum”. The elevation of the well’s depth reference point

with respect to the permanent datum. Usually defined either as the

top of the KB (Kelley bushing) or DF (drill floor). Typically, the

elevation of the top of the KB is identical to the elevation of the DF.

38 SRE (aka SREF) Source reference elevation. Airgun depths are measured with respect

to the current SRE. SRE is normally zero for a marine survey. For a

land survey over rough terrain, the elevation of ground level at one

source location may be quite different to the elevation of ground

level at another location. A pre-surveyed peg marking SRE provides

a convenient local reference against which a vibrator pad elevation

or an airgun depth, may be measured.

39 SD Source depth below SRE.

40 S2M Distance from source to source monitor transducer.

41 Source_elevation_error In a heaving sea, the actual source elevation may vary from shot to

shot. An integrating source motion sensor may be used to determine

the error in elevation for each shot.

42 External_reference_delay_us Typically the source controller radio delay, measured in

microseconds. The reference channel must be advanced by this

delay (transit time increases with positive delay).

43 Supplied_Ts_us Time correction to seismic datum, measured in microseconds.

44 TB_advance_us Delay from TB (electronic timebreak) to first recorded sample. This

field provided because TB may occur between sampling instants.

45 Tool_skew_us Timing skew correction from surface to digital downhole channels

in microseconds, as already applied internally by the acquisition

controller (diagnostic).

46 Raw_SCX_x10 The raw source easting coordinate as received in a positioning fix

string, no units conversion, but multiplied by 10 (diagnostic).

47 Raw_SCY_x10 The raw source northing coordinate as received in a positioning fix

string, no units conversion, but multiplied by 10 (diagnostic).

48 Controller_second Modulo N seconds tick count. Controller ticks provide short term

holdover timing continuity in the event of GPS signal loss.

49 Microsecond Elapsed microseconds from the last controller tick that occurred

before the first sample. Holds precision UTC microseconds if

timestamp_mode = 2.

50 Timestamp_mode Specifies source of the timestamp.

0 = local time from PC

1 = UTC time from PC

2 = UTC time from GPS (accurate to within 10us of UTC).

51 SDE Seismic datum elevation.


of 13

52 Error_control 0 = None

1 = Parity Control

53 Microseismic_mode 0 = Non-Microseismic Data

1 = Microseismic Data

54 Overlap_samples Number of samples to overlap in consecutive records (typically 0 for

non-microseismic files, 1 for microseismic)

55-128 Reserved

7 Channel Header Block

The channel header block contains a contiguous array of channel structures. The number of channel structures is

specified in the "channels_defined" variable in the general header block. The first channel structure contains

information for channel one, etc. Each channel structure comprises sixteen 32-bit little-endian fields (64 bytes

per channel). The 32-bit fields to be interpreted as either 32-bit signed integers or 32-bit IEEE754 floating point,

as indicated below.

Note that a channel structure will exist for all channels available to the instrument, but not all of these channels

will have been used. The NS field specifies the number of samples present for any particular channel, a value of

zero signalling that the channel was switched off and so has no data. Use the format_code to determine the

number of bytes required for each sample.


of 13

8 Channel structure

Field Type Attribute Description

1 Int32 Legacy_Owner This field maintained for compatibility with legacy MIRF

readers. New readers should use field 6 below. 0 =

Reference channel 1-98 = Receiver 1-98 99 = Auxiliary

channel

2 Int32 Descriptor 0 = no descriptor

1 = VZ (vertical component)

2 = HX (horizontal transverse component)

3 = HY (horizontal axial component)

4 = DH (downhole hydrophone)

5 = HYDRO (monitor hydrophone)

6 = PILOT (Vibroseis pilot sweep)

7 = TB (timebreak)

8 = GF (Vibroseis ground force)

9 = RM (Vibroseis reaction mass)

10 = BP (Vibroseis base-plate)

11 = PPS (GPS Pulse Per Second)

12 = VSIM (Vibrator similarity)

13 = RSIM (Radio similarity)

14 = NGEO (Near-field Geophone)

15 = FGEO (Far-field Geophone)

3 Int32 Format_code Data format for this channel

4 Int32 NS Number of samples recorded for this channel. Interpret

NS=0 to mean that this channel unused.

5 Int32 Pointer_us Cursor position in microseconds.

6 Int32 Owner -2 = AUX

-1 = REF

0 = none

1...N = Receiver 1... Receiver N

7 Int32 RCX Receiver easting coordinate relative to wellhead.

8 Int32 RCY Receiver northing coordinate relative to wellhead.

9 Int32 TVD True vertical depth below WRE.

10 Int32 MDO Measured depth offset from tool reference. Add to Tool_MD

to obtain the correct measured depth for this channel.

11 Float32 HSI See Section 13

12 Int32 Reserved

13 Float32 SSF Per-channel sensor scaling factor. Used to apply client-

defined polarity or amplitude correction. Note: this value is

(integer) zero for pre-MIRF4 files.

14 Float32 DC DC offset in volts.

15 Float32 SF Scaling factor to convert sample value to input voltage.

Note: this value is (integer) zero for original MIRF-1 files.

16 Float32 Max_magnitude Greatest sample magnitude present in this channel


of 13

9 Sample Encoding formats

A MIRF file contains a variably sized header block followed by contiguous de-multiplexed channel data blocks

arranged in channel order. The size of each channel block is determined by the number of samples recorded for

that channel multiplied by the number of bytes required to encode each sample. The number of samples

recorded for a given channel is specified by the NS field of the channel structure for that channel and may be

zero. The number of bytes required for each sample depends upon the format_code, which specifies how

samples are encoded.

To facilitate data transmission and minimise storage requirements, raw data is often encoded to sixteen bits (two

bytes) per sample using the instantaneous floating point (IFP) format native to the acquisition hardware. There is

currently only one IFP format in use: Geochain (also used by the unsupported SST500 and MSR systems).

There are also three obsolete IFP formats defined: DAQ, Multilock and Delta. The decoding algorithm is similar

for all IFP formats.

The GSP controller uses an efficient 24-bit integer format (three bytes) for surface channels. Two 32-bit (four

byte) encoding formats are also defined, 32-bit signed integer and 32-bit IEEE754 floating point. The floating

point format is always used for pre-processed data (e.g. correlated or stacked data).

A signed 16-bit integer format (2 bytes) is also used by certain downhole systems.

9.1 Format –1 (FORMAT_VARIES)

From MIRF version 3 onwards, each trace may have a different data format. If the format_code value in the

general header block equals –1 (0xffffffff), then use the per-channel format_code from the channel header block

to determine the correct data format.

9.2 Format 0 (DAQ IFP) (obsolete)

Format 0 uses sixteen bits (two bytes) per sample. The fourteen most significant bits hold the twos complement

mantissa and the two least significant bits hold the exponent. The exponent may have one of four possible

values, zero through three, corresponding to the four possible gains of x1, x8, x64 and x512.

9.3 Format 1 (Multilock IFP) (obsolete)

Format 1 uses sixteen bits (two bytes) per sample. The twelve most significant bits hold the twos complement

mantissa and the four least significant bits hold the exponent. The exponent may have one of twelve possible

values, zero through eleven, corresponding to the twelve possible instrument gains of x1, x2, x4, x8, x16, x32,

x64, x128, x256, x512, x1024 and x2048.

9.4 Format 2 (Geochain IFP, also used for SST500 & MSR)

Format 2 uses sixteen bits (two bytes) per sample. The fourteen most significant bits hold the twos complement

mantissa and the two least significant bits hold the exponent. The exponent may have one of four possible

values, zero through three, corresponding to the four possible instrument gains of x1, x4, x16 and x64.

The last MIRF file in a Continuous Recording session will be padded out with zeroes to ensure it is 1 seconds in

duration.


of 13

IFP Error handling

A consequence of the Geochain IFP coding algorithm is that the numbers 0, 1 & 2 are not valid codes. For

example, a sample value of zero would cause both exponent bits to be set (maximum instrument gain), yielding

a coded integer value 3. The integers 0, 1 & 2 are therefore reserved to indicate sample status as follows:

0 = no data received.

1 = detected telemetry error

2 = reserved.

Any sample flagged as a detected telemetry error, should normally be replaced by a value interpolated from

adjacent good samples. Some caution is required, if many contiguous errors are present, then the record may be

unacceptably distorted.

9.5 Format 3 (32-bit integer)

Format 3 uses thirty-two bits (four bytes) per sample, representing a standard 32-bit binary signed integer value.

This format is defined for possible future use.

9.6 Format 4 (32-bit floating point)

Format 4 uses thirty-two bits (four bytes) per sample, representing a little-endian IEEE754 floating point value,

the standard "float" type in the "C" language. This format is used for stack files generated by the ACQ

acquisition software and by AS-272 digitisers in formatted mode.

For AS-272 digitisers, the full-scale value is 0.0390625V. Error conditions are indicated by out-of-range values.

A value of 1.0 indicates a receiver missing sample, the equivalent of ‘0’, NO_DATA in IFP mode.

A value of 2.0 indicates a telemetry missing sample, the equivalent of ‘1’, TELEMETRY_ERROR in IFP mode.

9.7 Format 5 (24-bit Integer)

Format 5 uses twenty four bits (three bytes) per sample, representing a 24-bit binary signed integer value. This

format is big-endian, the first byte holds the eight most significant bits. This format is used for GSP surface

channels.

9.8 Format 6 (16-bit Integer)

Format 6 uses sixteen bits (two bytes) per sample, representing a standard little-endian 16-bit binary signed

integer value, the standard “short” type in the “C” language. This format is used by SMWD systems.

9.9 Format 7 (Delta IFP) (obsolete)

Format 7 uses sixteen bits (two bytes) per sample. The twelve least significant bits hold the twos complement

mantissa and the four most significant bits hold the exponent. The exponent may have one of twelve possible

values, zero through eleven, corresponding to the twelve possible instrument gains of x1, x2, x4, x8, x16, x32,

x64, x128, x256, x512, x1024 and x2048.


of 13

10 Sample decoding

The decoding process is very similar for all three IFP formats. The differences are the size of the mantissa and

the number and meaning of the exponent gain codes. The following explanation uses 32-bit little-endian "C"

language semantics. Avalon Sciences can supply a copy of the “C” source file which contains our own decoding

algorithms.

In the case of the Geochain format (format 2) the decoding procedure is as follows..

(1) Access the two-byte IFP sample as a 16-bit "short".

(2) Check that sample is not one of the special flag values 0,1 or 2.

(3) Use the two least significant bits to obtain the exponent integer (gain code, 0-3).

(4) Divide the sample integer by four (two right arithmetic bit-shifts) to obtain the mantissa integer.

(5) Convert the mantissa to floating-point form.

(6) Divide the mantissa by the IFP gain value obtained from the gain code exponent.

(7) Multiply the result by the per-channel scaling factor (SF), to recover instrument input terminal

voltage.

(8) Subtract the per-channel DC_volts to remove the measured dc offset.

(9) Multiply by the per-channel sensor_scaling_factor (SSF) to apply any sensor polarity or calibration

correction.

11 Legacy MIRF exceptions

The last three steps (scaling to input voltage, dc offset subtraction and sensor calibration) should be applied to

all formats. Note however that for pre-MIRF4 files, either or both of the SF or sensor_SF fields may contain

(integer) zero, so must be ignored. The recommended procedure is to check if the value is equal to zero and if so

then replace the zero with unity.

12 High Side Indication

This field (field 11) is only used for channels representing a three axis receiver that supports High Side

Indication. When used, this field represents a different attribute for each of the three channels (as determined by

the Channel Descriptor field – 2) in a receiver as follows:

Channel Descriptor Type Attribute Value

VZ - Vertical component Bool HSI Valid 0 = NOT valid, 1 = Valid

HX – Horizontal Transverse

component

Float32 HSI Inclination HSI Inclination Angle (-90 to +90 degrees)

HY – Horizontal Axial

component

Float32 HSI Roll HSI Roll Angle (-180 to +180 degrees)

avalon sciences limited td000795 · communicated to unauthorised persons or persons outside the...

Documents